Systems and Methods for Automatic Eye Gaze Refinement

1. A method implemented in a computing device having a front-facing camera, comprising: responsive to the front-facing camera capturing a digital image of an individual, applying facial landmark detection and identifying eye regions in the digital image; for at least one of the eye regions, performing the steps of: extracting attributes of the eye region; determining an eye gaze score based on the extracted attributes; generating a modified eye region based on the eye gaze score, wherein generating the modified eye region comprises adjusting at least one of: a direction of a pupil with respect to the front-facing camera, a size of the pupil, and a shape of the pupil; and outputting a modified digital image with the modified eye region.

2. The method of claim 1 , wherein determining the eye gaze score based on the extracted attributes comprises: identifying a lower eyelid and an upper eyelid; generating a first coefficient based on a curvature of the lower eyelid; generating a second coefficient based on a curvature of the upper eyelid; and determining the eye gaze score based on the first and second coefficients.

3. The method of claim 2 , wherein determining the eye gaze score based on the first and second coefficients comprises: retrieving weight values from a data store and generating weighted first and second coefficients; and determining the eye gaze score based on the weighted first and second coefficients.

4. The method of claim 2 , wherein generating the modified eye region comprises: determining a warping vector based on the second coefficient; and adjusting the curvature of the upper eyelid based on the warping vector.

5. A method implemented in a computing device having a front-facing camera, comprising: responsive to the front-facing camera capturing a digital image of an individual, applying facial landmark detection and identifying eye regions in the digital image; for at least one of the eye regions, performing the steps of: extracting attributes of the eye region; determining an eye gaze score based on the extracted attributes; generating a modified eye region based on the eye gaze score, wherein generating the modified eye region based on the eye gaze score comprises: determining a direction of an eye gaze with respect to the front-facing camera based on the eye gaze score; and warping a pupil in the eye region to modify the direction of the eye gaze while maintaining an original curvature of the pupil; and outputting a modified digital image with the modified eye region.

6. A system, comprising: a front-facing camera; a memory storing instructions; a processor coupled to the memory and configured by the instructions to at least: responsive to the front-facing camera capturing a digital image of an individual, apply facial landmark detection and identifying eye regions in the digital image; for at least one of the eye regions, the processor is further configured to: extract attributes of the eye region; determine an eye gaze score based on the extracted attributes; generate a modified eye region based on the eye gaze score, wherein the processor is configured to generate the modified eye region by adjusting at least one of: a direction of a pupil with respect to the front-facing camera, a size of the pupil, and a shape of the pupil; and output a modified digital image with the modified eye region.

7. The system of claim 6 , wherein the processor is configured to determine the eye gaze score based on the extracted attributes by: identifying a lower eyelid and an upper eyelid; generating a first coefficient based on a curvature of the lower eyelid; generating a second coefficient based on a curvature of the upper eyelid; and determining the eye gaze score based on the first and second coefficients.

8. The system of claim 7 , wherein the processor is configured to determine the eye gaze score based on the first and second coefficients by: retrieving weight values from a data store and generating weighted first and second coefficients; and determining the eye gaze score based on the weighted first and second coefficients.

9. The system of claim 7 , wherein the processor is configured to generate the modified eye region by: determining a warping vector based on the second coefficient; and adjusting the curvature of the upper eyelid based on the warping vector.

10. A system, comprising: a front-facing camera; a memory storing instructions; a processor coupled to the memory and configured by the instructions to at least: responsive to the front-facing camera capturing a digital image of an individual, apply facial landmark detection and identifying eye regions in the digital image; for at least one of the eye regions, the processor is further configured to: extract attributes of the eye region; determine an eye gaze score based on the extracted attributes; generate a modified eye region based on the eye gaze score, wherein the processor is configured to generate the modified eye region based on the eye gaze score by determining a direction of an eye gaze with respect to the front-facing camera based on the eye gaze score and warping a pupil in the eye region to modify the direction of the eye gaze while maintaining an original curvature of the pupil; and output a modified digital image with the modified eye region.

11. A non-transitory computer-readable storage medium storing instructions to be implemented by a computing device having a front-facing camera and a processor, wherein the instructions, when executed by the processor, cause the computing device to at least: responsive to the front-facing camera capturing a digital image of an individual, apply facial landmark detection and identifying eye regions in the digital image; for at least one of the eye regions, the processor is further configured to: extract attributes of the eye region; determine an eye gaze score based on the extracted attributes; generate a modified eye region based on the eye gaze score, wherein the processor is configured to generate the modified eye region by adjusting at least one of: a direction of a pupil with respect to the front-facing camera, a size of the pupil, and a shape of the pupil; and output a modified digital image with the modified eye region.

12. The non-transitory computer-readable storage medium of claim 11 , wherein the processor is configured to determine the eye gaze score based on the extracted attributes by: identifying a lower eyelid and an upper eyelid; generating a first coefficient based on a curvature of the lower eyelid; generating a second coefficient based on a curvature of the upper eyelid; and determining the eye gaze score based on the first and second coefficients.

13. The non-transitory computer-readable storage medium of claim 12 , wherein the processor is configured to determine the eye gaze score based on the first and second coefficients by: retrieving weight values from a data store and generating weighted first and second coefficients; and determining the eye gaze score based on the weighted first and second coefficients.

14. The non-transitory computer-readable storage medium of claim 12 , wherein the processor is configured to generate the modified eye region by: determining a warping vector based on the second coefficient; and adjusting the curvature of the upper eyelid based on the warping vector.

15. A non-transitory computer-readable storage medium storing instructions to be implemented by a computing device having a front-facing camera and a processor, wherein the instructions, when executed by the processor, cause the computing device to at least: responsive to the front-facing camera capturing a digital image of an individual, apply facial landmark detection and identifying eye regions in the digital image; for at least one of the eye regions, the processor is further configured to: extract attributes of the eye region; determine an eye gaze score based on the extracted attributes; generate a modified eye region based on the eye gaze score, wherein the processor is configured to generate the modified eye region based on the eye gaze score by determining a direction of an eye gaze with respect to the front-facing camera based on the eye gaze score and warping a pupil in the eye region to modify the direction of the eye gaze while maintaining an original curvature of the pupil; and output a modified digital image with the modified eye region.